This is a repository copy of Delayed meal timing after exercise is associated with reduced
appetite and energy intake in adolescents with obesity.
White Rose Research Online URL for this paper:
https://eprints.whiterose.ac.uk/168421/
Version: Accepted Version
Article:
Fillon, A, Beaulieu, K orcid.org/0000-0001-8926-6953, Miguet, M et al. (8 more authors)
(2020) Delayed meal timing after exercise is associated with reduced appetite and energy
intake in adolescents with obesity. Pediatric Obesity, 15 (9). e12651. ISSN 2047-6310
https://doi.org/10.1111/ijpo.12651
© 2020 World Obesity Federation. This is the peer reviewed version of the following
article: Fillon, A, Beaulieu, K, Miguet, M, et al. Delayed meal timing after exercise is
associated with reduced appetite and energy intake in adolescents with obesity. Pediatric
Obesity. 2020; 15:e12651. , which has been published in final form at
https://doi.org/10.1111/ijpo.12651. This article may be used for non-commercial purposes
in accordance with Wiley Terms and Conditions for Use of Self-Archived Versions.
eprints@whiterose.ac.uk https://eprints.whiterose.ac.uk/ Reuse
Items deposited in White Rose Research Online are protected by copyright, with all rights reserved unless indicated otherwise. They may be downloaded and/or printed for private study, or other acts as permitted by national copyright laws. The publisher or other rights holders may allow further reproduction and re-use of the full text version. This is indicated by the licence information on the White Rose Research Online record for the item.
Takedown
If you consider content in White Rose Research Online to be in breach of UK law, please notify us by
1
Delayed meal timing after exercise is associated with reduced appetite and
1energy intake in adolescents with obesity
23
Abstract
4
Background. While the beneficial effects of exercise on appetite might depend on its timing during
5
the day or relative to a meal, this remains poorly explored in youth. 6
Objectives. To examine the importance of meal timing (+30vs.+90minutes) after performing exercise
7
on energy intake, appetite and food reward in adolescents with obesity. 8
Methods. Eighteen adolescents with obesity randomly completed 3 conditions: i) lunch (12:00pm) set 9
30min after a rest session (11:00am); ii) lunch (12:00pm) set 30min after an exercise session (11:00 10
am)(MEAL-30); iii) lunch (01:00pm) set 90min after an exercise session (11:00am)(MEAL-90). Lunch 11
and dinner ad libitum energy intake was assessed, food reward (LFPQ) assessed before and after lunch, 12
and before dinner, appetite sensations were assessed at regular intervals. 13
Results. Energy intake was lower at MEAL-90 than MEAL-30 and CON at lunch (p<0.05 and p<0.01, 14
respectively) and lunch+dinner combined(p<0.001). A decrease in intake (g) of protein, fat and 15
carbohydrate was observed. Post-exercise hunger was lower on MEAL-90 compared with CON. No 16
condition effects were found at lunch for food reward. 17
Conclusions. Delaying the timing of the meal after exercise might help affect energy balance by 18
decreasing ad libitum energy intake without increasing hunger and by improving satiety in adolescents 19
with obesity. 20
21
Key words. Exercise Timing, Appetite, Energy Intake, Obesity, Adolescent, Food reward
22
Clinical Trial reference: NCT03968458
2 24
3
Introduction
25
While practitioners and clinicians constantly work on the improvement of their weight loss 26
interventions, trying to identify the best exercise characteristics (modality, intensity, duration, etc.) to 27
prescribe, the need to also consider the timing of exercise has been recently suggested 1. Recent
28
studies effectively show that the beneficial effects of exercise might also depend on its timing during 29
the day or its delay/position regarding a meal 1. Some studies for instance showed that performing
30
acute exercise one to three hours after a meal could enhance the glycemic response in patients with 31
type II diabetes 2–5 while others showed a better postprandial lipemia response when exercise was
32
performed immediately before the meal 6–8.
33
Looking at the alarming progression of overweight and obesity among children and adolescents, it 34
seems necessary to deepen our understanding on the effects of exercise on overall energy balance, in 35
order to optimize our weight loss strategies. It is now clear that physical exercise does not only impact 36
energy expenditure, it also affects energy intake and appetite control in youth and adolescents with 37
obesity 9. The current literature mainly investigated the effect of exercise duration 10,11, intensity 12–14
38
or modality 15 on subsequent food intake, appetite sensations or food reward, while the potential role
39
played by the timing of exercise remains poorly explored 16.
40
In 2017, Mathieu and collaborators assessed the effects of exercising immediately before or after a 41
lunch meal in primary school children on overall energy balance 17. Although they did not observe any
42
difference on energy intake between conditions (before or after the meal), their results highlight the 43
beneficial effect of performing pre-meal moderate-to-vigorous over low-intensity exercise on 44
subsequent energy intake 17. More recently, similar results were obtained among adolescents with
45
obesity whose energy intake and food reward remained unchanged whether the adolescents 46
performed 30 min of cycling exercise (65% VO2peak) immediately before or after their lunch meal 18.
47
Interestingly, others investigated the potential effect of the delay between an acute exercise bout and 48
the following meal on energy intake and appetite. In their work, Albert and colleagues compared the 49
effects of exercising (treadmill running at 70% VO2max) 45 min or 180 min before lunch, in normal
4 weight adolescents 19. The authors observed an 11% reduction of the adolescents’ ad libitum energy
51
intake and a 23% decrease in fat intake when the exercise was performed 45 min before lunch, 52
compared to 180 min. Moreover, there were no difference in terms of appetite sensations and no 53
energy compensation at the following snack or dinner. Our research group recently examined the 54
effect of the exercise-meal delay on energy intake, appetite and food reward among adolescents with 55
obesity 20. According to our results, a 30-min cycling exercise bout (65% VO
2max) performed 60 min
56
before lunch favored a 14% reduction of ad libitum energy intake while the same exercise performed 57
180 min before lunch did not affect the adolescents’ energy intake. While appetite sensations (hunger, 58
fullness, prospective food consumption and desire to eat) did not differ between conditions, our 59
results also showed a significantly lower pre-meal explicit liking for high-fat relative to low-fat foods 60
when the exercise was set close to the meal, suggesting the implication of the food reward system 20.
61
Altogether, these results seem to show a beneficial effect of exercising close to a meal on overall 62
energy balance in adolescents. 63
Although these studies compared exercises of similar characteristics (e.g. duration, modality, 64
intensity), their metabolic demand might have been different due to their divergent delay from 65
breakfast, which might have important implications when it comes to subsequent energy intake. 66
Indeed, it has been shown that the metabolic activity during exercise, particularly the contribution of 67
the energy substrates, is different depending on the delay between a breakfast and this exercise 21.
68
The substrate oxidation during exercise, especially the rate of carbohydrate oxidation has been 69
associated with subsequent energy intake 22, particularly in adults with obesity 23,24. Investigating the
70
effect of the timing of exercise on appetite and energy intake needs to consider not only its delay with 71
the following meal but also the time interval between exercise and the previous food intake. 72
In that context, the aim was to examine the importance of meal timing (+30 or +90 minutes) after 73
performing exercise on energy intake, appetite and food reward in adolescents with obesity. 74
75
Materials and methods
5
Participants
77
Eighteen adolescents with obesity (according to 25) aged 12-15 years (Tanner stage 3-4) were enrolled
78
in this study (12 boys (12.6 ± 1.2 years) and 6 girls (13.0 ± 1.6 years)). They were recruited through the 79
local Pediatric Obesity Center (Tza Nou, La Bourboule, France), based on the following main inclusion 80
criteria: i) to be free of any medication known to influence appetite or metabolism; ii) to be free of 81
any contraindication to physical activity; iii) to be classified as physically inactive (taking part in less 82
than 2 hours of physical activity per week as assessed using the International Physical Activity 83
Questionnaire –IPAQ 26). This study was conducted in accordance with the Helsinki declaration and all
84
the adolescents and their legal representative received information sheets and signed consent forms 85
as requested by the local ethical authorities (Human Ethical Committee authorization reference: 2019-86
A00530-57; Clinical Trial reference: NCT03968458). 87
1.1. Design
88
After a preliminary medical inclusion visit performed by a pediatrician to control for the ability of the 89
adolescents to complete the study, they were asked to perform a maximal aerobic test and their body 90
composition was assessed by dual-energy x-ray absorptiometry (DXA). The adolescents thereafter 91
completed the three following experimental sessions (one week apart) in randomized order: i) lunch 92
(at 12:00pm) set 30 min after a rest session (at 11:00 am) ii) lunch (at 12:00pm) set 30 min after an 93
exercise session (at 11:00am; MEAL-30); iii) lunch (at 1:00pm) set 90 min after an exercise session (at 94
11:00am; MEAL-90). On the three occasions, participants received a standardized breakfast (08:00am) 95
and were asked to remain at rest (CON) or to cycle for 30 min at 11:00am and eat either 30 min (on 96
MEAL-30; lunch at 12:00pm) or 90 min (on MEAL-90; lunch at 1:00pm) after exercise. Dinner was 97
provided to the adolescents at 6:30pm. They were asked to complete the Leeds Food Preference 98
Questionnaire (LFPQ) 27 before and after the lunch meal and before dinner. Lunch and dinner energy
99
intake were assessed via ad libitum buffet-style meals. Appetite sensations were measured at regular 100
intervals throughout the day. Outside the experimental conditions and between the two ad libitum 101
test meals, the adolescents stayed in the laboratory, devoid of any food cues, and were requested not 102
6 to engage in any moderate-to-vigorous physical activity and mainly completed sedentary activities 103
such as reading, homework or board games. Figure 1 details the whole design of the study. 104
………Figure 1……….. 105
1.2. Anthropometric characteristics and body composition
106
Body mass and height were measured wearing light clothing while bare-footed, using a digital scale 107
and a standard wall-mounted stadiometer, respectively. Body mass index (BMI) was calculated as 108
body mass (kg) divided by height squared (m²) and the sex and age dependent French reference curves 109
were used to obtain the BMI percentile 28. Fat mass (FM) and fat-free mass (FFM) were assessed by
110
dual-energy X-ray absorptiometry (DXA) following standardized procedures (QDR4500A scanner, 111
Hologic, Waltham, MA, USA). These measurements were obtained during the preliminary visit by a 112
trained technician. 113
1.3. Peak oxygen uptake test (V̇O2peak)
114
Each adolescent performed a V̇O2peak test on a traditional ergometer 29. The initial power was set at
115
30W during 3 minutes, followed by a 15W increment every minute until exhaustion. The adolescents 116
were strongly encouraged by the experimenters throughout the test to perform their maximal effort. 117
Maximal criteria were: heart rate >90% of the theoretical maximum heart rate (210 − 0.65 × age), 118
respiratory exchange ratio (RER = V̇CO2/V̇O2) > 1.1 and/or V̇O2 plateau. Cardiac electrical activity
119
(Ultima SeriesTM, Saint Paul, MN) and heart rate (Polar V800) were monitored and the test was 120
coupled with a measurement of breath-by-breath gas exchanges (BreezeSuite Software, Saint Paul, 121
MN), that determined V̇O2 and V̇CO2. Volumes and gases were calibrated before each test. V̇O2peak was
122
defined as the average of the last 30 s of exercise before exhaustion. 123
1.4. Experimental conditions
124
Rest condition (CON): During this condition, the adolescents were asked to remain quiet and were not
125
allowed to engage in any physical activity. They were asked to stay seated on a comfortable chair (30 126
7 min) between 11:00am and 11:30am, not being allowed to talk, read, watch TV or to complete any 127
intellectual tasks. Energy expenditure was assessed during the 30-min rest period using portable 128
indirect calorimetry (K4b², COSMED Inc., Rome, Italy). 129
Lunch condition 30 min after exercise (MEAL-30): Between 11:00am and 11:30am, the participants
130
performed a 30-min moderate-intensity exercise bout (65% VO2peak) on a cycle ergometer. The
131
intensity was controlled by heart rate records (Polar V800) using the results from the maximal aerobic 132
capacity testing. Exercise-induced energy expenditure was calculated based on the results obtained 133
during the maximal oxygen uptake test. 134
Lunch condition 90 min after exercise (MEAL-90): The adolescents performed the same exercise bout
135
as MEAL-30 and at the same time, but the ad libitum lunch meal was served at 1:00pm (90 min after 136
the end of the exercise). 137
1.5. Energy intake
138
At 08:00am, the adolescents consumed a standardized calibrated breakfast (500 kcal) respecting the 139
recommendations for their age (composition: bread (50 g), butter (10 g), marmalade (15 g), yoghurt 140
(125 g) or semi-skimmed milk (20 cl), fruit or fruit juice (20 cl)). Lunch and dinner meals were served 141
ad libitum using a buffet-type meal. The content of the buffets was determined using a food
142
preference and habits questionnaire filled in by the adolescents during the inclusion visit, as previously 143
described 30. Top rated items as well as disliked items and items liked but not usually consumed were
144
excluded to avoid over-, under- and occasional consumption. The lunch menu was beef steak, pasta, 145
mustard, cheese, yoghurt, compote, fruits and bread. The dinner menu was ham/turkey, beans, 146
mashed potato, cheese, yoghurt, compote, fruits and bread. Food items were presented in abundance 147
and the adolescents were told to eat until comfortably full. Adolescents made their choices and 148
composed their trays individually before joining their habitual table (5 adolescents per table). Lunch 149
and dinner were served in a quiet environment free of music, cellphones or television. Food items 150
were weighed by the experimenters before and after each meal. Energy intake and macronutrient 151
8 composition (proportion of fat, carbohydrate and protein) were calculated using the software Bilnut 152
4.0. This methodology has been previously validated and published 30. Lunch and total relative energy
153
intake (REI) were calculated such as: energy intake – exercise-induced energy expenditure. 154
1.6. Subjective appetite sensations
155
Appetite sensations were collected at regular intervals throughout the day using visual analogue scales 156
(150-mm scales) 31. Adolescents had to report their hunger, fullness, desire to eat (DTE) and
157
prospective food consumption (PFC) before and immediately after breakfast, prior and after rest 158
(CON) or exercise (MEAL-30 and MEAL-90), before and immediately after lunch, 30 min and 60 min 159
after lunch, before and immediately after dinner. 160
1.7. Food liking and wanting
161
The Leeds Food Preference Questionnaire, described in greater methodological detail by Dalton and 162
Finlayson 32, provided measures of food preference and food reward. The adolescents were presented
163
with a culturally (food items and language) adapted version of the LFPQ following the recent 164
recommendations from Oustric and collaborators 33. Participants were presented with an array of
165
pictures of individual food items common in the diet. Foods were chosen by the local research team 166
from a validated database to be either predominantly high (>50% energy) or low (<20% energy) in fat 167
but similar in familiarity, protein content, palatability and suitable for the study population. The LFPQ 168
has been deployed in a range of research 32 including a recent exercise/appetite trial in young French
169
males 34 and adolescents 20,35,36.
170
Explicit liking was measured by participants rating the extent to which they like each food (“How 171
pleasant would it be to taste this food now?”). The food images were presented individually, in a 172
randomized order and participants made their ratings using a 100-mm VAS. Implicit wanting was 173
assessed using a forced choice methodology in which the food images were paired so that every image 174
from each of the four food types was compared to every other type over 96 trials (food pairs). 175
Participants were instructed to respond as quickly and accurately as they could to indicate the food 176
9 they want to eat the most at that time (“Which food do you most want to eat now?”). Reaction times 177
for all responses were covertly recorded and used to compute mean response times for each food 178
type after adjusting for frequency of selection. 179
Responses on the LFPQ were used to compute mean scores for high-fat, low-fat, sweet or savoury 180
food types (and different fat-taste combinations). Fat bias scores were calculated as the difference 181
between the high-fat scores and the low-fat scores, with positive values indicating greater liking or 182
wanting for high-fat relative to low-fat foods and negative values indicating greater liking or wanting 183
for low-fat relative to high-fat foods. Sweet bias scores were calculated as the difference between the 184
sweet and savoury scores, with positive values indicating greater liking or wanting for sweet relative 185
to savoury foods and negative values indicating greater liking or wanting for savoury relative to sweet 186
foods. 187
1.8. Statistical analysis
188
Statistical analyses were performed using Stata software, Version 13 (StataCorp, College Station, TX, 189
US). The sample size estimation was determined according to (i) CONSORT 2010 statement, extension 190
to randomized pilot and feasibility trials (Eldridge et al. CONSORT 2010 statement: extension to 191
randomized pilot and feasibility trials. Pilot and Feasibility Studies (2016) 2:64) and (ii) Cohen’s 192
recommendations 37 who has defined effect-size bounds as : small (ES: 0.2), medium (ES: 0.5) and large
193
(ES: 0.8, “‘grossly perceptible and therefore large”). So, with 15 patients by condition, an effect-size 194
around 1 can be highlighted for a two-sided type I error at 1.7% (correction due to multiple 195
comparisons), a statistical power greater than 80% and an intra-class correlation coefficient at 0.5 to 196
take into account between and within participant variability. All tests were two-sided, with a Type I 197
error set at 0.05. Continuous data was expressed as mean ± standard deviation (SD) or median 198
[interquartile range] according to statistical distribution. The assumption of normality was assessed 199
by using the Shapiro-Wilk test. Daily (total) area under the curve (AUC) were calculated using the 200
trapezoidal method. Random-effects models for repeated data were performed to compare three 201
10 conditions (i) considering the following fixed effects: time, condition and time x condition interaction, 202
and (ii) taking into account between and within participant variability (subject as random-effect). A 203
Sidak’s type I error correction was applied to perform multiple comparisons. As proposed by some 204
statisticians 38,39 a particular focus will be also given to the magnitude of differences, in addition to
205
inferential statistical tests expressed using p-values. The normality of residuals from these models was 206
studied using the Shapiro-Wilk test. When appropriate, a logarithmic transformation was proposed to 207
achieve the normality of dependent outcome. 208
209
2. Results
210
Eighteen adolescents with obesity participated in this study. Their mean age was 12.7 ± 1.3 years, 211
body weight was 88.9 ± 23.6 kg (with a BMI of 33.3 ± 6.5 kg/m2 (z-BMI 2.2 ± 0.4), with a percentage of
212
body fat mass of 37.6 ± 5.0 % and a FFM of 53.1 ± 12.5 kg. 213
The adolescents had a V̇O2peak of 21.8 ± 4.6 ml/min/kg. Energy expenditure induced by the exercise
214
(total duration 30 min) was significantly higher compared to the 30-min resting energy expenditure 215
(168.8 ± 43.6 kcal and 46.9 ± 14.9 kcal, respectively; p<0.001). 216
Table 1 details the results related to absolute and relative energy intake. At lunch, absolute ad libitum 217
energy intake was significantly lower in MEAL-90 than MEAL-30 and CON (p<0.05 and p<0.01, 218
respectively) and in MEAL-30 than CON (p<0.05). Dinner ad libitum energy intake was significantly 219
lower in MEAL-90 compared with MEAL-30 (p<0.01) with no difference between the exercise 220
conditions and CON. Total daily absolute ad libitum energy intake was significantly lower in MEAL-90 221
compared with both CON and MEAL-30 (p<0.001). 222
REI at lunch was significantly higher in CON compared with MEAL-30 and MEAL-90 (p<0.05 and 223
p<0.001, respectively) and total REI was significantly higher in CON compared with MEAL-90 (p<0.001). 224
Both lunch (p<0.05) and total REI (p<0.001) were significantly lower in MEAL-90 than MEAL-30. 225
11 ………Table 1………. 226
The lunch and total absolute intake of protein, fat were significantly lower in MEAL-90 compared with 227
both CON (p<0.01 and p<0.05, respectively) and MEAL-30 (p<0.01 and p<0.05, respectively) while their 228
intake at dinner was significantly lower in MEAL-90 compared with MEAL-30 (p<0.05). The absolute 229
intake of CHO was significantly lower in MEAL-90 compared with CON at lunch (p<0.05) and 230
significantly higher in MEAL-30 compared with CON at dinner (p<0.05). Total absolute CHO intake was 231
only significantly lower in MEAL-90 compared with CON (p<0.05). No significant difference was 232
observed between conditions regarding the relative intake of each macronutrient. Table 2 details 233
these results. 234
………Table 2……… 235
Figure 2 presents the results related to appetite sensations. Fasting hunger, fullness, PFC and DTE did 236
not differ between conditions. After the standardized breakfast, significant differences between 237
conditions were found: hunger and DTE were higher in MEAL-30 than MEAL-90 (p=0.003 and p=0.02), 238
respectively) and CON (p=0.010 and p=0.016, respectively), while PFC was greater in MEAL-30 than 239
MEAL-90 only (p=0.021). Before exercise, hunger was significantly lower during both exercise 240
conditions than during CON (p<0.001 for both). After exercise, this difference remained significant 241
only between CON and MEAL-90 (p=0.004). Immediately before lunch, hunger and PFC were 242
significantly lower in MEAL-30 compared with CON (p=0.036 and p=0.041, respectively). Post-lunch 243
sensations were similar between conditions. Pre-dinner hunger was lower during both exercise 244
conditions compared with CON (p=0.006 for MEAL-30 and p=0.003 for MEAL-90). Pre-dinner fullness 245
was greater in MEAL-30 and MEAL-90 compared with CON (p=0.006 and p=0.003, respectively). 246
Regarding pre-dinner DTE and PFC, only MEAL-90 was significantly lower than CON (p=0.006 and 247
p=0.005, respectively). Concerning the daily AUC (Figure 2), relative to CON, hunger and DTE were 248
significantly lower in MEAL-30 (p=0.019 and p=0.05, respectively) and MEAL-90 (p=0.034 and p=0.031, 249
respectively). 250
12 ………Figure 2……….. 251
252
As detailed in Table 3, there was a significant condition effect for pre-dinner explicit liking fat bias 253
(p=0.004), with explicit liking for high-fat foods being lower in MEAL-90 compared with both CON 254
(p=0.001) and MEAL-30 (p=0.004). While explicit liking taste bias significantly decreased in response 255
to the lunch meal during the CON condition (p<0.001), this significant meal effect disappeared during 256
both exercise conditions, without a meal x condition interaction. Implicit wanting taste bias 257
significantly increased in response to the lunch test meal during MEAL-90 (p=0.04), and no meal effect 258
was observed in CON and MEAL-30. 259
……….Table 3……….. 260
Discussion
261
The timing of exercise relative to a meal has been recently highlighted for its influence on energy 262
intake and appetite control 1,16, with some recent studies suggesting a better effect of acute exercise
263
performed close to a meal on energy intake and appetite in both adolescents who are lean 19 and
264
adolescents with obesity 20. However these studies did not consider the potential impact of the delay
265
between the exercise and the previous breakfast intake. It has been shown that this delay will impact 266
the metabolic nature of exercise such as the substrates used 21, which might, in turn, differently affect
267
subsequent energy intake 22–24. In that context, the aim of the present study was to investigate the
268
effect of exercise performed at the same delay from breakfast on energy intake, appetite sensations 269
and food reward at the following lunch set either 30 or 90 min after exercise in adolescents with 270
obesity. 271
According to our results, both exercise conditions (MEAL-30 and MEAL-90) led to significantly lower 272
absolute energy intake at lunch compare to CON. This is in line with previous studies in similar 273
populations showing reduced subsequent intake in response to acute exercise set at the same time of 274
the morning 12,14,20,40. Interestingly, absolute energy intake was also significantly lower in MEAL-90
275
compared with MEAL-30, suggesting a greater anorexigenic effect when exercise does not 276
13 immediately precede the meal. Additionally, total and dinner absolute energy intake were lower 277
during MEAL-90 only, with total daily energy intake reduced by 12% (250 kcal/day) and 16% (352 278
kcal/day) compared with CON and MEAL-30, respectively. These results are reinforced by a lower 279
lunch relative energy intake after MEAL-30 compared with CON and lower lunch and total REI during 280
MEAL-90 compared with both MEAL-30 and CON. Importantly, while most of the available evidence 281
supports the anorexigenic effect of intensive exercise 13,35,41,42, our results reinforce more recent work
282
also observing reduced food intake in response to moderate-to-vigorous exercise in adolescents and 283
children with obesity 40,40.
284
While available evidence indicates the beneficial effect of exercising close to a meal on subsequent 285
energy intake 19,20, our results seem to suggest that more than the exercise-meal delay itself, the
286
interval between the exercise and the following eating episode is of importance. 287
A balanced buffet meal offering several items selected to avoid any over-, under- or occasional-288
consumption (as previously validated 30) was offered to adolescents which provided the opportunity
289
to also assess their macronutrient intake. While none of the relative intake of fat, protein and 290
carbohydrate were found different between conditions, their absolute consumption at lunch was 291
reduced only in MEAL-90 compared with CON, and compared with MEAL-30 for protein and lipid. 292
Interestingly, the absolute intake of carbohydrate at dinner increased in MEAL-30 compared with the 293
two other conditions. The macronutrient responses observed in MEAL-90 seem in line with Albert et 294
al. in lean adolescents 19 and with our previous study in adolescents with obesity 20, showing reduced
295
absolute macronutrient intake after moderate exercise set at the end of the morning. The current 296
study however missed to find similar results in MEAL-30, suggesting here the potential importance of 297
the delay between the exercise and the previous eating episode (breakfast). Indeed, in these previous 298
studies, the appetitive responses to exercise set at different times of the morning, and then at 299
different delays from breakfast, were compared, meaning that despite similar duration, modality and 300
intensity, the exercise was not of similar metabolic and energetic load 21, which might explain our
14 results. Unfortunately, it was not possible in the present study to measure the substrate oxidation 302
during exercise and at rest. Furthermore, it remains difficult to reach a consensus regarding the effect 303
of acute exercise on macronutrient intake in lean adolescents and in adolescents with obesity based 304
on the available evidence 42.
305
Regarding the adolescents’ subjective appetite sensations, our results show a lower daily (AUC) 306
hunger and desire to eat in both exercise conditions compared with CON. Although pre-lunch hunger 307
and PFC were significantly lower in MEAL-30 compared with CON, which could have contributed to 308
the lower observed ad libitum energy intake, they remained unchanged in MEAL-90 while the 309
decreased food consumption was even more pronounced. This inconsistency between appetite 310
sensations and energy intake reinforce the previously described uncoupling effect of exercise between 311
these sensations and food consumption 43. Interestingly however, post-lunch sensations were
312
identical between exercise conditions, suggesting a similar satiating effect of lunch meals despite 313
lower intakes in MEAL-30 and particularly in MEAL-90, limiting any potential subsequent 314
compensatory responses. This is even reinforced by the significantly reduced food intake observed at 315
dinner in MEAL-90. This is of particular importance since energy deficits, especially when induced by 316
reduced energy intake, have been shown to generate a subsequent compensatory rise in food intake, 317
with physical exercise limiting or avoiding such a compensation 34,44.
318
Some recent studies have highlighted the importance of considering the effect of exercise on food 319
reward to better understand its impact on subsequent energy intake in adolescents with obesity 35.
320
We also assessed whether the liking and wanting for food could be impacted by the delay between 321
eating episodes and exercise in this population. In 2018, Miguet and colleagues observed reduced 322
relative preference for fat and sweet taste, and implicit wanting for high-fat foods (also using the 323
LFPQ) in response to an ad libitum meal set 30 minutes after a 16-minute cycling high intensity interval 324
exercise in a similar population 35. According to the present results, none of the pre or post lunch
325
components of liking and wanting were different between conditions. These results are contradictive 326
15 with those from Miguet et al. (2018), especially regarding our MEAL-30 condition that had the same 327
delay between the exercise and the meal. However, the exercise intensities were different (high 328
intensity intermittent exercise vs. moderate intensity continuous exercise), reinforcing once more the 329
importance of the exercise intensity in the subsequent control of energy intake. Interestingly, we can 330
see here a significantly lower explicit liking for high-fat food immediately before dinner in MEAL-90 331
compared with the two others, which might contribute to the observed reduced dinner ad libitum 332
food intake. Our results are however also in contradiction with some recently published from our 333
group, showing different food reward responses depending on exercise-meal timing in adolescents 334
with obesity 20. A lower pre-meal explicit liking for high-fat relative to low-fat foods was observed
335
when the adolescents performed 30 min of moderate intensity cycling 60 min before lunch compared 336
with the same exercise performed 180 min before lunch 20. The different LFPQ timing between
MEAL-337
90 and the two other conditions must be considered when interpreting our results. Indeed, food 338
reward was assessed pre- and post- lunch meaning that its delay from exercise was different, which 339
might have affected the results. Although there is a growing interest in the effect of exercise on food 340
reward in this population, evidence remains too limited to draw any conclusion and further studies 341
using standardized designs are needed. 342
The present results must be interpreted in light of some limitations. First, as for the other published 343
studies examining the timing of exercise relative to a meal 16,17,19,20, the lack of direct evaluation of the
344
adolescents’ oxygen consumption and substrate oxidation using indirect calorimeters, as well as the 345
lack of a lean control group to examine the potential weight status effect, are the two main limitations. 346
Although the laboratory-based nature of this work constitutes a strength as it allows a better control 347
of the adolescents’ activity and intake, it might also not be representative of their habitual daily free-348
living setting, such as the school setting for instance, as previously underlined by Mathieu et al. in 349
healthy adolescents 17. Finally, the lack of tracking of the adolescents’ food intake over 24 to 48 hours
350
for practical reasons also limits the interpretation of our results 12.
16 In line with the present work, another potential important factor, while not addressed in the current 352
study, is the timing of exercise (and food intake) with regards to circadian/diurnal rhythms. Emerging 353
evidence suggests that the timing of exercise 45,46 (and food intake 47,48) impact body weight regulation.
354
Any effects observed from exercise-meal delays may be a result of an interaction with 355
circadian/diurnal oscillations occurring relative to sleep/wake times. Future studies should propose a 356
more complete and integrative exploration of the chronobiologic regulations of energy intake and 357
overall energy metabolism in such adolescents with obesity. Indeed, not only the timings of exercise 358
and /or energy intake should be considered, but also their interactions with the adolescents’ sleep, to 359
better understand and potentially regulate their 24-hour circadian rhythm 49,50. Some key physiological
360
actors of this circadian clock, such as ghrelin and leptin for instance, who are particularly involved in 361
the control of appetite and respondents to sleep and exercise should be mainly considered 51.
362 363
Conclusion
364
To conclude, the present study reinforces the interest in the timing of exercise relative to a meal to 365
affect overall energy balance in youth with obesity; highlighting the importance of the time interval 366
between both the exercise and the previous eating episode, and the exercise and the following meal. 367
According to these results, delaying the timing of the meal after exercise might help reduce energy 368
balance by decreasing ad libitum energy intake without increasing hunger and by improving satiety in 369
adolescents with obesity. Future studies should question the importance of the exercise-meal timing 370
on the longer term. While further acute and chronic studies are needed, these results contribute to 371
the current limited body of evidence in the area and seem important in order to optimize weight loss 372
strategies. 373
374
Conflicts of interest statement
375
None and this research did not receive any specific grant from funding agencies in the public, 376
commercial, or not-for-profit sectors. 377
17 378
379
Author contributions 380
AF and DT conceived experiments. AF, MM and MB carried out experiments, AF and DT analysed data. 381
KB was involved in writing the paper and all authors had final approval of the submitted and published 382 versions. 383 384 Acknowledgements 385
The authors are grateful to all of the adolescents that participated in the program, and to the Nutrition 386
Obesity Ambulatory Hospital (UGECAM) that provided their generous support. 387
18
Table legends
389
Table 1: Absolute and Relative Energy Intake in response the three conditions. 390
Table 2: Macronutrient Intake in response the three conditions. 391
Table 3: Pre- and Post-test meal food reward on the three experimental conditions 392
393 394
Figures legends
395
Figure 1 : Study design 396
Figure 2 : Daily appetite sensations and AUC for hunger, fullness, prospective food consumption and 397
desire to eat 398
19
Table 1: Absolute and Relative Energy Intake in response the three conditions.
400
CON: control condition; MEAL-30: Test meal 30 min after exercise; MEAL-90: Test meal 90 min after exercise; SD: Standard 401
Deviation; ES: Effect Size; *p<0.05 vs. CON ; **p<0.01 vs. CON ; ***p<0.001 vs. CON ; ap<0.05 MEAL-30 vs. MEAL-90 ; 402
bp<0.01 MEAL-30 vs. MEAL-90 ; cp<0.001 MEAL-30 vs. MEAL-90; ES: post hoc effect size 403
404
CON MEAL-30 MEAL-90
p
ES
Mean (SD) Mean (SD) Mean (SD) CON vs. MEAL-30 CON vs. MEAL-90
MEAL-30 vs. MEAL-90 E n e rgy Int ak e (k cal ) Lunch 1380 (185) 1347 (313)* 1168 (234)**a 0.0143 -0.12[-0.60, 0.35] -0.71[-1.19, -0.24] 0.59[0.11, 1.06] Dinner 796 (294) 931 (260) 748 (245)b 0.0363 0.48[0.00, 0.96] -0.20[-0.67, 0.28] 0.68[0.20, 1.15] Total 2175 (330) 2277 (476) 1925 (360)***c 0.0001 0.27[-0.21, 0.74] -0.80[-1.28, -0.33] 1.07[0.59, 1.54] R e lat iv e E ne rg y Int ak e ( k cal ) Lunch 1337 (188) 1172 (313)* 1006 (246)***a 0.0003 -0.56[-1.03, -0.08] -1.08[-1.56, -0.61] 0.52[0.04, 1.00] Total 2119 (332) 2110 (489) 1755 (366)***c <0.0001 -0.11[-0.58, 0.37] -1.16[-1.63, -0.68] 1.06[0.59, 1.54]
20
Table 2: Macronutrient Intake in response the three conditions.
405
CON MEAL-30 MEAL-90
p
ES
Mean (SD) Mean (SD) Mean (SD)
CON vs. MEAL-30 CON vs. MEAL-90 MEAL-30 vs.
MEAL-90 P rot e ins (g) Lunch 73.8 (11.5) 71.9 (17.2) 60.7 (13.9)**b 0.0059 -0.13[-0.61, 0.34] -0.76[-1.24, -0.29] 0.63[0.15, 1.10] Dinner 42.0 (18.4) 46.8 (14.4) 37.2 (13.2)a 0.1811 0.25[-0.22, 0.73] -0.30[-0.78, 0.17] 0.56[0.08, 1.03] Total 115.9 (22.6) 118.7 (23.8) 98.8 (19.4)***c 0.0007 0.08[-0.40, 0.55] -0.85[-1.32, -0.37] 0.93[0.45, 1.40] P rot e ins (% ) Lunch 21.5 (2.3) 21.4 (3.0) 20.8 (2.3) 0.5108 0.05[-0.42, 0.53] -0.07[-0.55, 0.40] 0.23[-0.25, 0.70] Dinner 20.8 (5.2) 19.9 (3.1) 20.1 (3.6) 0.8811 0.17[-0.31, 0.64] 0.01[-0.46, 0.49] -0.06[-0.53, 0.42] Total 21.3 (2.5) 21.0 (2.0) 20.6 (2.3) 0.6248 0.10[-0.38, 0.58] -0.05[-0.53, 0.42] 0.14[-0.33, 0.62] Li pi ds (g ) Lunch 45.4 (9.6) 45.0 (14.2) 38.1 (12.5)*a 0.0146 -0.06[-0.53, 0.42] -0.54[-1.01, -0.06] 0.48[0.06, 1.01] Dinner 28.8 (19.0) 33.8 (15.1) 26.1 (14.3)a 0.0642 0.33[-0.15, 0.80] -0.18[-0.66, 0.30] 0.51[0.03, 0.98] Total 74.3 (18.0) 78.8 (19.9) 65.8 (19.1)*b 0.0123 0.25[-0.23, 0.72] -0.54[-1.01, -0.06] 0.79[0.31, 1.26] Li pi ds (% ) Lunch 29.8 (5.8) 30.3 (8.0) 29.2 (7.3) 0.1910 0.05[-0.42, 0.53] -0.07[-0.55, 0.40] 0.13[-0.35, 0.60] Dinner 30.0 (12.9) 31.3 (10.6) 29.7 (9.8) 0.0277 0.17[-0.31 0.64] 0.01[-0.46, 0.49] 0.15[-0.32, 0.63] Total 30.7 (5.8) 31.2 (4.8) 30.5 (5.7) 0.9655 0.10[-0.38, 0.58] -0.05[-0.53, 0.42] 0.15[-0.32, 0.63] C HO ( g) Lunch 166.7 (39.4) 160.8 (52.8) 144.2 (34.6)* 0.1649 -0.14[-0.62, 0.33] -0.52[-0.99, -0.04] 0.37[-0.10, 0.85] Dinner 92.8 (31.5) 109.9 (31.5)* 91.9 (29.4)a 0.0269 0.52[0.04, 0.99] -0.036[-0.54, 0.41] 0.58[0.11, 1.06] Total 259.5 (56.1) 270.7 (70.0) 233.9 (49.7)a 0.0751 0.17[-0.31, 0.64] -0.45[-0.92, 0.03] 0.61[0.14, 1.09] C HO ( % ) Lunch 48.0 (7.6) 47.5 (10.5) 49.5 (9.1) 0.2149 0.06[-0.53, 0.42] 0.15[-0.33, 0.62] -0.20[-0.68, 0.27] Dinner 49.7 (15.6) 48.9 (12.4) 50.7 (10.7) 0.0840 -0.01[-0.48, 0.47] 0.13[-0.34, 0.61] -0.14[-0.61, 0.34] Total 47.8 (7.4) 47.4 (6.1) 48.7 (7.3) 0.9547 -0.05[-0.53, 0.42] 0.14[-0.34, 0.61] -0.19[-0.67, 0.28]
CON: control condition; MEAL-30: Test meal 30 minutes after exercise; MEAL-90: Test meal 90 minutes after exercise; SD: 406
Standard Deviation; *p<0.05 vs. CON ; **p<0.01 vs. CON ; ***p<0.001 vs. CON ; ap<0.05 MEAL-30 vs. MEAL-90 ; bp<0.01 407
MEAL-30 vs. MEAL-90 ; cp<0.001 MEAL-30 vs. MEAL-90; ES: Effect Size; CHO: Carbohydrates; ES: post hoc effect size. 408
21
Table 3: Pre- and Post-test meal food reward on the three experimental conditions
410
CON MEAL-30 MEAL-90
p Interaction time x condition
Mean (SD) Mean (SD) Mean (SD) CON vs. MEAL-30 CON vs. MEAL-90 MEAL-30 vs. MEAL-90 Implicit Wanting Fat Bias Before lunch 22.32 (31.15) 19.96 (33.15) 22.80 (31.68) 0.78 0.99 0.58 0.56 After lunch 20.21 (45.58) 17.63 (48.49) 12.61 (29.50) 0.46 p before vs. after lunch 0.88 0.80 0.90 0.00[-0.48-0.48] -0.13[-0.61-0.34] -0.14[-0.62-0.33] Before dinner 4.37 (64.45) 20.74 (19.89) 14.99 (26.63) 0.49 Taste Bias Before lunch 31.60 (33.67) 34.17 (41.81) 24.90 (32.49) 0.76 0.93 0.14 0.26 After lunch 25.60 (54.02) 27.00 (67.00) 43.59 (30.79) 0.59 p before vs. after lunch 0.69 0.85 0.04 0.02[-0.45-0.50] 0.36[-0.11-0.84] 0.27[-0.20-0.75] Before dinner 38.24 (37.81) 40.40 (40.11) 42.30 (28.12) 0.98 Explicit Liking Fat Bias Before lunch 10.02 (19.71) 12.52 (16.35) 10.53 (19.64) 0.34 0.57 0.77 0.86 After lunch 5.29 (9.39) 5.14 (10.66) 4.08 (9.25) 0.94 p before vs. after lunch 0.27 0.03 0.11 -0.14[-0.61-0.34] -0.07[-0.55-0.40] 0.04[-0.43-0.52] Before dinner 11.35 (19.83) 9.04 (16.34) 2.44 (13.00)***b <0.001 Taste Bias Before lunch 26.18 (20.37) 21.95 (23.03) 20.31 (22.89) 0.82 0.10 0.25 0.74 After lunch 12.78 (19.10) 18.08 (25.78) 14.47 (27.62) 0.73 p before vs. after lunch <0.001 0.38 0.19 0.40[-0.07-0.88] 0.28[-0.19-0.76] -0.08[-0.56-0.40] Before dinner 24.00 (24.58) 21.40 (26.08) 20.76 (28.74) 0.99
CON: control condition; MEAL-30: Test meal 30 min after exercise; MEAL-90: Test meal 90 min after exercise; SD: 411
Standard Deviation; ***p<0.001 vs. CON; bp<0.01 MEAL-30 vs. MEAL-90 ; P values and Effect Size are presented for 412
interactions. 413
22 1. Reid RER, Thivel D, Mathieu M-E. Understanding the potential contribution of a third “T” to FITT 415
exercise prescription: the case of timing in exercise for obesity and cardiometabolic 416
management in children. Appl Physiol Nutr Metab. 2019;44(8):911-914. doi:10.1139/apnm-417
2018-0462 418
2. Borror A, Zieff G, Battaglini C, Stoner L. The Effects of Postprandial Exercise on Glucose Control 419
in Individuals with Type 2 Diabetes: A Systematic Review. Sports Med Auckl NZ. 2018;48(6):1479-420
1491. doi:10.1007/s40279-018-0864-x 421
3. Chacko E. Exercising Tactically for Taming Postmeal Glucose Surges. Scientifica. 422
2016;2016:4045717. doi:10.1155/2016/4045717 423
4. Haxhi J, Scotto di Palumbo A, Sacchetti M. Exercising for Metabolic Control: Is Timing Important. 424
Ann Nutr Metab. 2013;62(1):14-25. doi:10.1159/000343788
425
5. Teo SYM, Kanaley JA, Guelfi KJ, et al. Exercise Timing in Type 2 Diabetes Mellitus: A Systematic 426
Review. Med Sci Sports Exerc. 2018;50(12):2387-2397. doi:10.1249/MSS.0000000000001732 427
6. Petitt DS, Cureton KJ. Effects of prior exercise on postprandial lipemia: A quantitative review. 428
Metabolism. 2003;52(4):418-424. doi:10.1053/meta.2003.50071
429
7. Zhang JQ, Ji LL, Nunez G, Feathers S, Hart CL, Yao WX. Effect of exercise timing on postprandial 430
lipemia in hypertriglyceridemic men. Can J Appl Physiol Rev Can Physiol Appl. 2004;29(5):590-431
603. 432
8. Zhang JQ, Thomas TR, Ball SD. Effect of exercise timing on postprandial lipemia and HDL 433
cholesterol subfractions. J Appl Physiol. 1998;85(4):1516-1522. 434
doi:10.1152/jappl.1998.85.4.1516 435
9. Thivel D, Finlayson G, Blundell JE. Homeostatic and neurocognitive control of energy intake in 436
response to exercise in pediatric obesity: a psychobiological framework. Obes Rev Off J Int Assoc 437
Study Obes. 2019;20(2):316-324. doi:10.1111/obr.12782
438
10. Masurier J, Mathieu M-E, Fearnbach SN, et al. Effect of Exercise Duration on Subsequent 439
Appetite and Energy Intake in Obese Adolescent Girls. Int J Sport Nutr Exerc Metab. August 2018. 440
doi:10.1123/ijsnem.2017-0352 441
11. Tamam S, Bellissimo N, Patel BP, Thomas SG, Anderson GH. Overweight and obese boys reduce 442
food intake in response to a glucose drink but fail to increase intake in response to exercise of 443
short duration. Appl Physiol Nutr Metab. 2012;37(3):520-529. doi:10.1139/h2012-038 444
12. Thivel D, Isacco L, Montaurier C, Boirie Y, Duché P, Morio B. The 24-h Energy Intake of Obese 445
Adolescents Is Spontaneously Reduced after Intensive Exercise: A Randomized Controlled Trial 446
in Calorimetric Chambers. PLoS ONE. 2012;7(1). doi:10.1371/journal.pone.0029840 447
13. Thivel D, Isacco L, Rousset S, Boirie Y, Morio B, Duché P. Intensive exercise: A remedy for 448
childhood obesity? Physiol Behav. 2011;102(2):132-136. doi:10.1016/j.physbeh.2010.10.011 449
14. Thivel D, Metz L, Julien A, Morio B, Duché P. Obese but not lean adolescents spontaneously 450
decrease energy intake after intensive exercise. Physiol Behav. 2014;123:41-46. 451
doi:10.1016/j.physbeh.2013.09.018 452
23 15. Laan DJ, Leidy HJ, Lim E, Campbell WW. Effects and reproducibility of aerobic and resistance 453
exercise on appetite and energy intake in young, physically active adults. Appl Physiol Nutr 454
Metab Physiol Appl Nutr Metab. 2010;35(6):842-847. doi:10.1139/H10-072
455
16. Fillon A, Mathieu ME, Boirie Y, Thivel D. Appetite control and exercise: Does the timing of 456
exercise play a role? Physiol Behav. 2020;218:112733. doi:10.1016/j.physbeh.2019.112733 457
17. Mathieu M-E, Lebkowski A, Laplante E, Drapeau V, Thivel D. Optimal timing of exercise for 458
influencing energy intake in children during school lunch. Appetite. 2018;120:416-422. 459
doi:10.1016/j.appet.2017.09.011 460
18. Fillon A, Miguet M, Bailly M, et al. Does exercising before or after a meal optimize overall energy 461
balance in adolescents with obesity? In: Katowice: Annals of Nutrition and Metabolism.; 2019. 462
19. Albert M-H, Drapeau V, Mathieu M-E. Timing of moderate-to-vigorous exercise and its impact 463
on subsequent energy intake in young males. Physiol Behav. 2015;151:557-562. 464
doi:10.1016/j.physbeh.2015.08.030 465
20. Fillon A, Mathieu M-E, Masurier J, et al. Effect of exercise-meal timing on energy intake, appetite 466
and food reward in adolescents with obesity: the TIMEX study. Appetite. 2020;146:104506. 467
21. Aucouturier J, Isacco L, Thivel D, et al. Effect of time interval between food intake and exercise 468
on substrate oxidation during exercise in obese and lean children. Clin Nutr Edinb Scotl. 469
2011;30(6):780-785. doi:10.1016/j.clnu.2011.03.011 470
22. Hopkins M, Jeukendrup A, King NA, Blundell JE. The Relationship between Substrate Metabolism, 471
Exercise and Appetite Control: Does Glycogen Availability Influence the Motivation to Eat, 472
Energy Intake or Food Choice? Sports Med. 2011;41(6):507-521. doi:10.2165/11588780-473
000000000-00000 474
23. Burton FL, Malkova D, Caslake MJ, Gill JMR. Substrate metabolism, appetite and feeding 475
behaviour under low and high energy turnover conditions in overweight women. Br J Nutr. 476
2010;104(8):1249-1259. doi:10.1017/S0007114510002023 477
24. Hopkins M, Blundell JE, King NA. Individual variability in compensatory eating following acute 478
exercise in overweight and obese women. Br J Sports Med. 2014;48(20):1472-1476. 479
doi:10.1136/bjsports-2012-091721 480
25. Cole TJ, Bellizzi MC, Flegal KM, Dietz WH. Establishing a standard definition for child overweight 481
and obesity worldwide: international survey. BMJ. 2000;320(7244):1240. 482
26. Craig CL, Marshall AL, Sjöström M, et al. International physical activity questionnaire: 12-country 483
reliability and validity. Med Sci Sports Exerc. 2003;35(8):1381-1395. 484
doi:10.1249/01.MSS.0000078924.61453.FB 485
27. Finlayson G, King N, Blundell J. The role of implicit wanting in relation to explicit liking and 486
wanting for food: implications for appetite control. Appetite. 2008;50(1):120-127. 487
doi:10.1016/j.appet.2007.06.007 488
28. WHO Multicentre Growth Reference Study Group. WHO Child Growth Standards based on 489
length/height, weight and age. Acta Paediatr Oslo Nor 1992 Suppl. 2006;450:76-85. 490
24 29. Rowland TW. Does peak VO2 reflect VO2max in children?: evidence from supramaximal testing. 491
Med Sci Sports Exerc. 1993;25(6):689-693.
492
30. Thivel D, Genin PM, Mathieu M-E, Pereira B, Metz L. Reproducibility of an in-laboratory test meal 493
to assess ad libitum energy intake in adolescents with obesity. Appetite. 2016;105:129-133. 494
doi:10.1016/j.appet.2016.05.028 495
31. Flint A, Raben A, Blundell JE, Astrup A. Reproducibility, power and validity of visual analogue 496
scales in assessment of appetite sensations in single test meal studies. Int J Obes Relat Metab 497
Disord J Int Assoc Study Obes. 2000;24(1):38-48.
498
32. Dalton M, Finlayson G. Psychobiological examination of liking and wanting for fat and sweet 499
taste in trait binge eating females. Physiol Behav. 2014;136:128-134. 500
doi:10.1016/j.physbeh.2014.03.019 501
33. Oustric P, Thivel D, Dalton M, et al. Measuring food preference and reward: application and 502
cross-cultural adaptation of the Leeds Food Preference Questionnaire in human experimental 503
research.Oustric P, Thivel D, Dalton M, Beaulieu K, Gibbons C, Hopkins M, Blundell J, Finlayson 504
G.. Food Qual Pref. 2020; 80: 103824. doi: 10.1016/j.foodqual.2019.103824. Food Qual Pref. 505
2020;80. doi:doi: 10.1016/j.foodqual.2019.103824 506
34. Thivel D, Finlayson G, Miguet M, et al. Energy depletion by 24-h fast leads to compensatory 507
appetite responses compared with matched energy depletion by exercise in healthy young 508
males. Br J Nutr. 2018;120(5):583-592. doi:10.1017/S0007114518001873 509
35. Miguet M, Fillon A, Khammassi M, et al. Appetite, energy intake and food reward responses to 510
an acute High Intensity Interval Exercise in adolescents with obesity. Physiol Behav. 2018;195:90-511
97. doi:10.1016/j.physbeh.2018.07.018 512
36. Thivel D, Roche J, Miguet M, et al. Post- moderate intensity exercise energy replacement does 513
not reduce subsequent appetite and energy intake in adolescents with obesity. Br J Nutr. in 514
press. 515
37. Cohen J. Statistical Power Analysis for the Behavioral Sciences (2nd Ed.). Lawrence Erlbaum. New 516
Jersey; 1988. 517
38. Feise RJ. Do multiple outcome measures require p-value adjustment? BMC Med Res Methodol. 518
2002;2:8. doi:10.1186/1471-2288-2-8 519
39. Rothman K, Greenland S. Modern Epidemiology. 2nd edn. Philadelphia: Lippencott-Raven; 1998. 520
40. Fearnbach SN, Masterson TD, Schlechter HA, et al. Impact of imposed exercise on energy intake 521
in children at risk for overweight. Nutr J. 2016;15(1):92. doi:10.1186/s12937-016-0206-5 522
41. Prado WL, Lofrano-Prado MC, Oyama LM, et al. Effect of a 12-Week Low vs. High Intensity 523
Aerobic Exercise Training on Appetite-Regulating Hormones in Obese Adolescents: A 524
Randomized Exercise Intervention Study. Pediatr Exerc Sci. 2015;27(4):510-517. 525
doi:10.1123/pes.2015-0018 526
42. Thivel D, Rumbold PL, King NA, Pereira B, Blundell JE, Mathieu M-E. Acute post-exercise energy 527
and macronutrient intake in lean and obese youth: a systematic review and meta-analysis. Int J 528
Obes 2005. 2016;40(10):1469-1479. doi:10.1038/ijo.2016.122
25 43. Thivel D, Chaput J-P. Are Post-Exercise Appetite Sensations and Energy Intake Coupled in 530
Children and Adolescents? Sports Med. 2014;44(6):735-741. doi:10.1007/s40279-014-0160-3 531
44. Thivel D, Doucet E, Julian V, Cardenoux C, Boirie Y, Duclos M. Nutritional compensation to 532
exercise- vs. diet-induced acute energy deficit in adolescents with obesity. Physiol Behav. 533
2017;176:159-164. doi:10.1016/j.physbeh.2016.10.022 534
45. Alizadeh Z, Younespour S, Rajabian Tabesh M, Haghravan S. Comparison between the effect of 535
6 weeks of morning or evening aerobic exercise on appetite and anthropometric indices: a 536
randomized controlled trial. Clin Obes. 2017;7(3):157-165. doi:10.1111/cob.12187 537
46. Willis EA, Creasy SA, Honas JJ, Melanson EL, Donnelly JE. The effects of exercise session timing 538
on weight loss and components of energy balance: midwest exercise trial 2. Int J Obes 2005. July 539
2019. doi:10.1038/s41366-019-0409-x 540
47. Ruddick-Collins LC, Johnston JD, Morgan PJ, Johnstone AM. The Big Breakfast Study: Chrono-541
nutrition influence on energy expenditure and bodyweight. Nutr Bull. 2018;43(2):174-183. 542
doi:DOI: 10.1111/nbu.12323 543
48. Johnston JD. Physiological Responses to Food Intake Throughout the Day. Nutr Res Rev. 544
2014;27(1):107-118. doi:DOI: 10.1017/S0954422414000055 545
49. Summa KC, Turek FW. Chronobiology and Obesity: Interactions Between Circadian Rhythms and 546
Energy Regulation. Adv Nutr. 2014;5(3):312-319. 547
50. de Castro MA, Riccioppo Garcez M, Lopes Pereira J, Mara Fisberg R. Eating Behaviours and 548
Dietary Intake Associations With Self-Reported Sleep Duration of Free-Living Brazilian Adults. 549
Appetite. 2019;137:207-217.
550
51. Westerterp-Plantenga MS. Sleep, Circadian Rhythm and Body Weight: Parallel Developments. 551
Proc Nutr Soc. 2016;75(4):431-439. doi:DOI: 10.1017/S0029665116000227
552 553